Allows python NaN values to be written as null in the JSON data
#196
Conversation
…ll` in the JSON data see influxdata#195 Adds a new dependency to module `simplejson` but the same effect is difficult to achieve with base module `json` see http://stackoverflow.com/questions/28639953/python-nan-json-encoder
|
I can see the improvement. However, I don't really like adding simplejson as a dependency. Are there any other solutions? |
|
Yes, I understand adding a new dependency is not desirable. |
|
Hi guys, May I suggest an alternative solution to this problem? If the dataframe value is a list of floats, all NaN values in that list will be replaced with None thus making the numpy array object JSON compliant. The nice thing about this is that all of it happens before the serialization takes place. See one possible implementation here: 0feadd9. If you're not happy with it, we can make adjustments, but I think it's better not to play around with the JSON encoders. Cheers! |
|
Looks good to me. On 17 June 2015 at 06:45, Georgi Dimitrov notifications@github.com wrote:
|
|
Closed in favor of #204 |
JSON dump modified to allows python
NaNvalues to be written asnullin the JSON datasee #195
Adds a new dependency to module
simplejsonbut the same effect is difficult to achieve with base modulejsonsee http://stackoverflow.com/questions/28639953/python-nan-json-encoder